Towards Optimization of Tuned Mass Dampers in Suspended Bridges by Accounting for Non-Linear Behaviour

Abstract

This master thesis investigates the dynamic behaviour of suspended pedestrian bridges, with a focus on the non-linearities resulting from the quadratic dependency of cables stiffness to displacement, using the Victor-Neels bridge as a case study. A detailed numerical model was developed in the software Finelg, and was validated by an experimental campaign involving pedestrian induced vibrations. The results from the modal identification using COV-SSI method and the tension measurement in hangers provided meaningful information to refine the numerical model to replicate the real-life behaviour of the bridge as closely as possible.

Non-linear dynamic analyses, performed with sine sweep of varying amplitude, revealed a significant increase in structural stiffness beyond a certain excitation threshold, resulting in reduced displacement amplitude compared to the linear predictions as well as a frequency shift in the resonant frequency peak. Even though a linear analysis is conservative, it can lead to excessive displacement estimations, potentially resulting in unnecessary tuned-mass damper (TMD) installations or structural oversizing.

This work investigates the effectiveness of linear TMDs in linear and non-linear analyses. While TMDs tuned to the frequency resulting from the modal analysis were effective, their efficiency was increased by taking into account the frequency shift from the non-linear dynamic analyses. The results also showed that a comparable damping efficiency could be achieved with a lower-mass TMD, offering interesting perspective in terms of possible optimization. Other perspectives include the implementation of a non-linear TMD with an exponential damping force depending on the speed, to enhance performance to reduce the amplitudes of the peak response.